[patch] Move negative checks to util.c, from acpi.c (resend)

[David Zeuthen]

On Thu, 2005-08-18 at 00:32 +0200, Danny Kukawka wrote:
> On Wednesday 17 August 2005 23:56, Richard Hughes wrote:
> > Moves the negative checks to util.c so that pmu (and soon to be apm) can
> > use the same logic.
> >
> > Okay to commit?
> 
> Looks good, I can't see any problems with this patch at the first look. 

Looks OK to me too..

> Something other I already discussed with David and which maybe effect yout 
> patch:
> 
> I would prefer to set the remainig time and percentage to -1 if there is any 
> error during calculation to know if the battery is really empty or if there 
> was a error. With the current behavior we can't differ between this two 
> cases. 
> 
> A case to explain:
>    - a error during calculation (e.g. one value is unknow 
>      in /proc/acpi/battery/*) and the calculation of remainig time/percentage 
>      fails.
>    - HAL set remainig time/percentage to 0
>    - you have a rule in your powermanagement program (e.g. powersave or 
>      g-p-m ) "shutdown if remainig time == 0"
> 
> In this case you shutdown maybe the system already if the battery is full. But 
> if you get -1 as returnvalue you can control other values to get more fuzzy 
> information about remainig time/percentage or pop-up a dialog to inform the 
> user about this error.
> 
> Any ideas?

You raise some good points..

Actually, I'd rather we set the percentage to 100% and remaining time to
some high number too (either a number made up or derived from design
capacity etc.) in order to let e.g. gnome-power-manager et. al. not do
any crazy actions like shutting down.. 

I'm not sure we want -1 because that will make it more difficult to
implement gnome-power-manager et. al. and, really, the main point is
that the bug with ACPI is something that should be fixed.. In other
words, I don't think we should design for broken hardware... I'm not
super religious about this - what does other people think?

Cheers,
David